RFR: JDK-8298448: UndefinedBehaviorSanitizer [v2]

[Justin King]

On Mon, 12 Dec 2022 04:34:07 GMT, David Holmes <dholmes at openjdk.org> wrote:

>>> I think it requires much broader discussion as to whether OpenJDK is actively seen to endorse these tools. Why these tools? What if there are other tools, should we support them all?
>>> 
>>> I'm not saying use of these tools may not be useful, but actually incorporating them into OpenJDK is a decision that needs to be made at a higher-level IMO.
>> 
>> The sanitizers are integrated directly with GCC and Clang/LLVM and are used by projects such as the Linux kernel. They are also used by companies such as Facebook and Google, which IIRC maintain some of the largest closed source mono repositories on the planet. As the sanitizers are integrated directly with GCC and Clang/LLVM, they are extremely easy to use (no external dependencies), fast, and have no direct alternatives. An alternative would also need to be integrated with the compilers in order to be at par.
>> 
>> Additionally configuration options for using ASan already exist in OpenJDK, so that ship has kinda sailed.
>> 
>> If we feel strongly about a discussion, we should probably discuss all the sanitizers as a whole. However that discussion can be done in parallel, as ASan is already used. Just adding the options to OpenJDK does not mean it is endorsed.
>
> @jcking this is not ready for integration. You have one review from build team. You have no reviews from core-libs for launcher change. You haven't even bothered to address the comments I made on the actual changes.

@dholmes-ora Also etiquette-wise, is it preferred that the commenter resolve the conversation or the author, and the commenter re-open if they feel it is not resolved? I am used to latter workflow, but OpenJDK might have different expectations and I'd like to follow them.

-------------

PR: https://git.openjdk.org/jdk/pull/11604